Re: Netra X1 IDE controller/disk issues

[<rjs%fdy2.demon.co.uk@localhost>]

Michael <macallan%netbsd.org@localhost> wrote:
>On May 11, 2010, at 4:20 PM, Chris Ross wrote:
>
>>  After replacing the primary internal disk on my Netra X1, it's  
>> running 5.1_RC1 stably.  I've since added another disk to it, a  
>> Samsung SP2514N (ATA/133 250GB).  The first time after I boot, I can  
>> label that disk, and when I try to newfs it: (partition a is all of  
>> the disk except the first cylinder) with the following:
>>
>> # newfs -i $[32*1024] -O 2 /dev/rwd1a
>>
>>  It causes the kernel to get "lost interrupt" issues, eventually  
>> downgrading it to PIO.  I see the following on the console [before  
>> newfs even begins to print the status dots]:
>>
>> aceride0:1:0: lost interrupt
>>      type: ata tc_bcount: 512 tc_skip: 0
>> aceride0:1:0: bus-master DMA error: missing interrupt, status=0x21
>> wd1: transfer error, downgrading to Ultra-DMA mode 3
>> wd1a: DMA error writing fsbn 488396159 (wd1 bn 488397167; cn 484520  
>> tn 15 sn 62), retrying
>> aceride0:1:0: lost interrupt
>>      type: ata tc_bcount: 512 tc_skip: 0
>> aceride0:1:0: bus-master DMA error: missing interrupt, status=0x21
>
>I'd try a different, known good cable before anything else.

I see the same kinds of error on NetBSD/cats if I don't force the
controller to PIO mode in the kernel config.

There is a comment in the sparc64 GENERIC config that suggests the
same change is needed for Blade 100 and Netra X1 systems.

Robert Swindells